This version is made available in accordance with publisher policies. Please, cite

only the publisher version using the citation below:

Roberto Rosal, Antonio Rodríguez, José Antonio

Perdigón-Melón, Alice Petre, Eloy García-Calvo, María

José Gómez, Ana Agüera, Amadeo R. Fernández-Alba,

Degradation of caffeine and identification of the

transformation products generated by ozonation,

Chemosphere, Volume 74, Issue 6, February 2009, Pages

825-831, ISSN 0045-6535,

10.1016/j.chemosphere.2008.10.010.

Link to official URL: http://www.sciencedirect.com/science/article/pii/S0045653508012800

COVER PAGE

Degradation of caffeine

and identification of the

transformation products

generated by ozonation

Chemosphere, 74, 825-831, 2009

Chemosphere, 74, 825-831, 2009

Degradation of caffeine and identification of the transformation products

generated by ozonation

Roberto Rosal*1, Antonio Rodríguez1, José Antonio Perdigón-Melón1, Alice Petre1, Eloy García-Calvo1,

María José Gómez2, Ana Agüera2 and Amadeo R. Fernández-Alba2

1 Department of Chemical Engineering, University of Alcalá, 28771 Alcalá de Henares, Spain. 2 Department of Analytical Chemistry, University of Almería, 04010 Almería, Spain.

* Corresponding author: roberto.rosal@uah.es

Abstract

The ozonation of caffeine in water was performed at different pH values, including acidic conditions. Kinetic

experiments were conducted by adding pulses of a concentrated caffeine solution to ozone saturated water. The results

showed a rapid decrease of ozone concentration during the first 15 s after injection, followed by a gradual decline at a

much slower rate. The data were fitted to a second order kinetic model with rate constants increasing from 0.25 to 1.05

M-1s-1 for pH in the 3-10 range. The initial ozone consumption per mole of ozonated caffeine was greater at high pH

values, reflecting a higher ozone decomposition rate. The decomposition of ozone was positively affected by the

concentration of caffeine, an effect that could be attributed to the presence of a reaction intermediate from the

ozonation of caffeine that behaved as a strong promoter of ozone decomposition. A study of the transformation

products identified by LC-TOF-MS was carried out, which permitted a tentative degradation pathway to be proposed

persistent by-products to be identified at both pH 3 and 8. Most transformation products were the result of the opening

of the imidazole ring after breaking caffeine’s N7C8 double bond.

Keywords: Caffeine; ozonation; mineralization; oxidation intermediates

1. Introduction

The growing scarcity of water resources is one of the

most critical environmental problems that need to be

solved in the near future. The use of reclaimed water has

been highlighted as playing an important role in the

management of water resources. Generalized reuse of

wastewater will reduce the consumption of drinking

water and allow water pollution to be controlled in rivers

and groundwaters, especially in arid and semi-arid

regions which suffer periodic acute water shortages.

Reclaimed wastewater may find several applications such

as, for example, the irrigation of agricultural and urban

areas, cooling water for industrial plants and recharging

groundwater bodies (Petalaa et al., 2006). A major hurdle

to this endeavour is the evidence that the conventional

treatment of municipal wastewaters discharges many

biologically-active compounds into the environment. The

improvement of analytical techniques has enabled the

detection of a number of compounds not routinely

analysed in the past, not only in the effluent of sewage

treatment plants (STP), but in many surface streams and

groundwaters. Most of them belong to the group of

emerging pollutants, which are candidates for future

regulation based on their potential effects on human

health and on the environment (Gagné et al., 2006).

Ternes et al. (1998) identified over thirty drugs in the

influent of German Municipal Wastewater Treatment

Plants. In their wake, many studies have revealed the

presence of drugs, disinfectants and other active

substances in the environment (Kümmerer, 2001; Tauxe-

Wuersch et al., 2005). The need for new treatment

strategies is made more acute by the fact that many of

these pollutants escape to conventional wastewater

treatments with the risk of becoming ubiquitous (Carballa

et al., 2004).

Caffeine is by far, the most commonly used legal drug in

the world either in the form of beverages or in

combination with analgesics to enhance their effect.

Caffeine is widely metabolized by humans, the disposal

of the unconsumed coffee and caffeinated soft drinks

being the predominant source of caffeine introduced into

the wastewater treatment system (Tang-Liu et al., 1983).

As a result, caffeine has been detected in many surface

streams and STP effluents (Franke et al., 1995; Ternes et

al., 2001; Kolpin et al., 2002). Lee and Rasmussen

(2006) reported the presence of caffeine in samples

collected upstream of STP discharges, as well as

persistent concentrations downstream. This distribution

pattern, different from other organic pollutants, is

probably linked to the existence of additional sources

such as unauthorized discharges. Caffeine was also

detected in the effluent of the municipal STP located in

Alcalá de Henares at a level of 600 ng/L (Rodríguez et

al., in press). Once discharged, caffeine persists in the

environment because of its high solubility in water

(37.47±1.01 g/L according to Sriamornsak and Kennedy,

2007) and low octanol-water partition coefficient

(Gossset et al., 1983). These facts, together with the lack

Chemosphere, 74, 825-831, 2009

of biogenic sources, led to caffeine being proposed as a

tracer of domestic pollutants (Seiler et al., 1999).

Ozone has been used extensively for disinfection or to

remove dissolved organic matter from surface water,

groundwater and wastewater. When used in alkaline

conditions, the decomposition of ozone produces a

considerable amount of hydroxyl radicals and falls into

the group of Advanced Oxidation Processes. AOP which

make use of ozone are effective in the mineralization of

organic compounds with a substantial lowering of the

COD or TOC of many effluents.(Rodríguez et al., in

press). However, in some cases the formation of toxic

oxidation by-products have been reported (Fernández-

Alba et al., 2002). Carr and Baird (2000) studied the

ozonation of caffeine at pH 3 with or without the addition

of hydrogen peroxide. They found a bimodal

mineralization pattern that which they attributed to the

formation of partially oxidized intermediates that

undergo further oxidation. Kolonko et al. (1979), Stadler

et al. (1996), Telo et al., (1997) and Dalmazio et al.,

(2005) also studied the oxidation of caffeine and

proposed some reaction intermediates for the hydroxyl

mediated reaction. This work evaluates the ability of

ozonation to oxidize caffeine at different pH values

including acidic conditions at which the concentration of

hydroxyl radicals is low, explores the kinetics of ozone-

caffeine reaction and identifies some reaction

intermediates.

2. Experimental

2.1. Materials and ozonation procedure

Caffeine was supplied by Sigma-Aldrich (99% purity).

Reaction runs were carried out in a 5-L glass jacketed

reactor whose temperature was controlled by a Huber

Polystat cc2 thermostatic regulator. The temperature of

the liquid inside the reactor was monitored throughout

the experiment by means of a Pt100 RTD. Ozone was

produced by a corona discharge ozonator (Ozomatic,

SWO100) fed by oxygen (about 95% purity) produced by

an AirSep AS-12 PSA oxygen generation unit. A mixture

of ozone and oxygen was bubbled into the liquid by

means of a porous glass disk with a gas flow of 0.20

Nm3/h. At a prescribed time, a pulse of concentrated

caffeine was introduced inside the reactor and the gas

flow was stopped. Some additional experiments were

performed with a continuous gas flow throughout the run.

The reaction vessel was agitated at a rate of 500 rpm

using a four-blade turbine. Details are given elsewhere

(Rosal et al., in press). The initial ozone-to-caffeine

molar ration varied in a 0.6-8 range. The experiments

were carried out at pH in the 3-10 range, controlled by

pumping a diluted sodium hydroxide using a feed-back

PID control. During the run, certain samples were

withdrawn at prescribed intervals. Ozone was quenched

in samples by adding a concentrated solution of sodium

thiosulfate or by bubbling nitrogen.

2.2. Analysis

The concentration of dissolved ozone was determined by

means of a Rosemount 499A OZ amperometric analyser

equipped with Pt 100 RTD temperature compensation

and calibrated against the Indigo Colorimetric Method

(SM 4500-O3 B). The signal was transmitted via a

Rosemount 1055 SoluComp II Analyser to a Agilent

34970 Data Acquisition Unit and recorded. The

concentration of ozone in the gas phase was determined

with a non-dispersive UV Photometer Anseros Ozomat

GM6000 Pro calibrated and tested against a chemical

method. The temperature inside the reactor was measured

with a Pt100 RTD. pH was measured with a CRISON

electrode connected to a Eutech αlpha-pH100 PID

control device that delivered a solution of sodium

hydroxide by means of a LC10AS Shimadzu

chromatography pump. The system allowed pH to be

controlled within ±0.1 units throughout the experiment.

The signals from temperature and pH sensors were also

sent to the Data Acquisition Unit and recorded in a

computer. Total organic carbon (TOC) analyses were

carried out by means of a Shimadzu TOC-VCSH total

carbon organic analyzer equipped with an ASI-V

autosampler.

The identification of caffeine transformation products

was carried out by liquid chromatography coupled to

time-of-flight mass spectrometry (LC-TOF-MS). The

Agilent Series 1100 liquid chromatograph (Agilent

Technologies, Palo Alto, USA) was equipped with a

reversed-phase C18 analytical column of 3 mm x 250

mm, 5 um particle size (ZORBAX, SB-C18, Agilent

Technologies). The mobile phase was a mixture of 0.1%

formic acid and 5% Milli-Q water in acetonitrile as

mobile phase A and 0.1% formic acid in water as mobile

phase B (pH 3.5) at a flow rate of 0.4 mL min-1. A linear

gradient progressed from 10%A (initial conditions) to

100%A in 30 min, and then remained steady at 100%A

for 5 min. The injection volume was 20 µL. This HPLC

system was connected to a TOF mass spectrometer

(Agilent Technologies) equipped with an electrospray

interface operating under the following conditions:

capillary 4000 V, nebulizer 40 psig, drying gas 9 L min-1,

gas temperature, 300ºC; skimmer voltage, 60 V; octapole

dc1, 37.5 V; octapole rf, 250 V. Routine analyses of

caffeine were performed by HPLC using a Hewlett

Packard apparatus equipped with a Kromasil C18 250 x

4.6 mm column. The mobile phase was a mixture of

acetonitrile and water (75:25) adjusted to pH 2.5 using

orthophosphoric acid with an isocratic flow of 0.5

mL/min at room temperature. Injections were performed

with a 20 µL loop and UV detection was carried out at

273 nm.

3. Results and discussion

3.1. Kinetics of ozonation

The kinetic data corresponding to the ozonation of

caffeine were taken in runs in which the ozone gas flow

was stopped immediately after the injection of caffeine.

Chemosphere, 74, 825-831, 2009

Fig. 1 shows, the evolution of the concentration of

caffeine and ozone over time for runs performed at pH 3-

10. The concentration of dissolved ozone was monitored

with a sampling period of 5 s throughout the run and in

all cases showed a two-stage pattern consisting of an

initial rapid decline followed by a second period with a

lower reaction rate. The difference in reaction between

both periods was more pronounced for runs performed at

lower pH and can be associated with the rapid depletion

of caffeine that took place during the first part of the

runs. Caffeine reacted quickly within the initial 15 s

period before subsequently declining at a much slower

rate, a pattern that suggested the existence of an

intermediate compound whose oxidation was the rate

controlling step. Table 1 shows some results of the

ozonation of caffeine for runs performed at different pH

and initial ozone-to-caffeine molar ratio. The rate of

ozone depletion was in all cases greater for runs

performed at higher pH values, as expected due to the

role of hydroxide anion in the ozone decomposition

reaction (Tomiyasu et al., 1985; Hoigné, 1998). Thus, the

moles of ozone consumed per mole of caffeine ozonated

during the first 15 s increased from 0.5 to 3.8 by

increasing pH from 3 to 10. The conversion of caffeine

during the first 15 s decreased in parallel from 0.61 to

0.11 for increasing pH, a result that was probably linked

to the reduction of ozone exposure originated by the

increased ozone decomposition rate under alkaline

conditions.

The ozonation of a given compound in water is a

consequence of the combination of direct and indirect or

radical oxidation reactions. Assuming second order rate

expressions for the direct reaction of ozone with the

organic compound, its rate of disappearance is:

3 3

AO A O HO A HO

d Ck C C k C C

d t• •− = + [1]

The kinetic constants 3O

k and HO

k • represent the direct

and indirect ozone reactions respectively. The integration

of Eq. 1 requires information about the concentration of

hydroxyl radical that is generally so low that it becomes

very difficult to measures directly. Elovitz and coworkers

(Elovitz and von Gunten, 1999; Elovitz et al., 2000)

proposed a kinetic approach based on a parameter Rct

defined as the relationship between the integral ct-

exposure to ozone and to the hydroxyl radical, the two

oxidant species involved in the system:

3

HO

ct

O

C dtR

C dt

•

=∫∫

[2]

The combination of Eqs. 1 and 2 yields, for the ozonation

of a given organic compound with initial concentration

CAo, the following expression in which the independent

variable is integral ozone exposure:

( )3 3 3 3 30 0 0 0

lnt t t t

AoO O O ct O R OHO HO HO

A

Ck C dt k C dt k R k C dt k C dt

C• • •= + = + =∫ ∫ ∫ ∫

[3]

Fig. 2 shows the logarithmic decline of caffeine (A) as a

function of integral ozone exposure for runs performed at

pH 3-10 and with two different initial ozone-to-caffeine

molar ratios. During the first period (< 15 s), a rapid

reaction took place in all runs followed by a much slower

decline in caffeine concentration. The second slow

reaction was first order in the concentration of dissolved

ozone. The fitting of experimental results to Eq. 3

permitted kinetic constants to be calculated for the

ozonation reaction, kR. The results shown in Table 1

indicated that the ozonation rate was higher in alkaline

conditions due to the greater concentration of hydroxyl

radicals associated with the reaction of ozone with

hydroxide anion (Beltrán, 2004). On the other hand, the

initial conversion of caffeine was a function of pH and

decreased from 60% (pH 3) to about 10% (pH 10) as pH

increased (Table 1).

The reduction of TOC was low irrespective of the pH of

runs. Fig. 3 shows the profiles of caffeine, ozone and

TOC during an ozonation in which the ozone gas flow

was maintained throughout the run. In this case, caffeine

became depleted at about 30 min, a period in which the

concentration of dissolved ozone was low, and increased

steadily thereafter. For this run, the extent of

mineralization was less than 10% after 120 min, but it

remained low even for runs performed under alkaline

conditions. For example, TOC decline was 17% at 120

min in a run performed at pH 8 under constant ozone gas

flow. In runs where gas flow was stopped, the reduction

of TOC was always less than 4% in 120 min. These

results pointed to a rapid initial direct reaction between

caffeine and molecular ozone, probably an attack on the

N7=C8 double bond whose products kept the

concentration of dissolved ozone low as long as caffeine

existed in solution. The reaction rate of caffeine was low

even at pH 10 at which the concentration of hydroxyl

radicals should be higher.

The decomposition of ozone after the injection of

caffeine exhibited first-order kinetics in two stages. A

quick initial reaction was followed by a second period

with a lower rate constant, the transition being clearer for

runs performed at low pH. The transition was associated

with the end of the initial rapid consumption of caffeine,

suggesting that the decomposition of ozone was the

consequence of two parallel processes, one of which

involves a reactive intermediate produced from caffeine:

,

ProductsR R I

k kA I�������

Assuming the steady-state hypothesis, the concentration

of this intermediate must be proportional to that of

caffeine and the mass balance to ozone in solution yields:

3

3 3

,1

,0

,

O d R

d O O A

R I

d C k kk C C C

d t k− = + [4]

Chemosphere, 74, 825-831, 2009

The reaction constant kd,0 is a pseudo-first order rate

constant associated with the matrix effect. The

combination of the differential form of Eq. 3 and Eq. 4,

yields a functional form for the concentration of ozone as

a function of time:

33

3 3

,1

,0

,

R Ok C dtO d R

d O O Ao

R I

d C k kk C C C e

d t k

− ∫− = + [5]

At the beginning of the run, for low reaction time, the

exponential term still approached unity, so that the

apparent first-order constant for the decomposition of

ozone should be linear with the initial concentration of

caffeine:

,1

,0

,

( 0)d R

d d Ao

R I

k kk t k C

k→ = + [6]

Fig. 4 shows the apparent rate constant kd as a function of

the initial concentration of caffeine. Results reasonably

followed Eq. 6 for pH in the 3-8 range with an intercept

of 0.03 s-1 representing the matrix effect on the

decomposition of ozone free of reactive intermediats

produced from caffeine. The slope was 0.75 L mmol-1 s-1,

a low value which probably reflects a high reaction

constant kR,I.

3.2. Identification of ozonation products

The TOC measurements performed during the ozonation

experiments indicated a low extent of mineralization

even in conditions at which caffeine became completely

depleted (Fig. 3). The organic compounds originated

during the ozonation of caffeine largely remained in

solution and, consequently, the process is likely to

originate persistent organic intermediates which are

necessarily less oxidizable than caffeine. There is some

information available on reaction products from the

advanced oxidation of caffeine in different conditions.

Dalmazio et al. (2005) reported reaction intermediates of

caffeine under the oxidative conditions of the H2O2/UV,

TiO2/UV, and Fenton systems. They state that the initial

attack of hydroxyl radicals on the C4=C5 double bond

resulted in the formation of N,N-dimethylparabanic acid

further oxidized to bis(N-hydroxymethyl)parabanic acid.

It has also been reported that the oxidation of caffeine by

hydroxyl radicals yielded 1,3,7-trimethyluric acid, from

hydroxylation on C8, and theophyline and other N-

demethylated products, such as 1,3-dimethyluric acid

(Telo et al., 1997). Stadler et al. (1996) identified 6-

amino-5-(N-formylmethyl-amino)-1,3-dimethyl-uracil, a

product resulting from the opening of the imidazole ring.

In this work, the identification of caffeine oxidation by-

products was based on the accurate mass measurements

recorded by a time-of-flight mass spectrometric

instrument. Such measurements permit elemental

compositions to be proposed for the molecular ions of

the detected compounds as well as of their characteristic

fragment ions and sodium adducts, thus providing a high

grade of confidence in the structure assignation. Table 2

shows the measured and calculated mass of the

protonated ions, the error between them and the proposed

empirical formula corresponding to the compounds

identified by LC-TOF-MS. Accurate mass measurements

recorded for the protonated caffeine molecule (m/z

195.0875, C8H11N4O2) showed excellent agreement

between experimental and calculated m/z values with an

error lower than 0.8 ppm. From the experiments of

caffeine ozonation at pH 3, four compounds were

identified as possible transformation products (see Figure

5a). The compound marked as P1 in Fig. 5 presented an

accurate mass of 243.0723, which corresponded to a

mass increase of 48 Da with respect to the protonated

caffeine molecule. The empirical formula that best

matches this experimental mass was C8H10N4O5 with an

error of 0.42 ppm. Therefore, this compound is likely to

be the product of the ring opening of a primary

molozonide (P0) resulting from the incorporation of

ozone to the N7=C8 double bond of the purine structure.

Molozonides are unstable and their ready decomposition

leads, in the case of carbon-carbon double bonds, to a

carbonyl fragment and a zwitterionic species leading in

water to the corresponding carboxyl compound (Cataldo,

2001). The structure proposed for this transformation

product was confirmed by the presence of the

corresponding sodium adduct at m/z 265.0547

(C8H10N4O5Na, 1.35 ppm error) and the fragment ion at

m/z 225.0620 Da (C8H8N4O4, 0.75 ppm error), identified

as corresponding to the neutral loss of H2O, which is

consistent with the presence of a COOH group and the

increase in DBE (double bond equivalency). The loss of

oxygen and an increase of two hydrogen atoms, probably

by reduction of the N=O group of P1, led to P2, which

was confirmed by the presence of the sodium adduct at

accurate mass of 251.0750 Da (C8H12N4O4Na, 0.09 ppm)

and by the reduction of the DBE, as a consequence of the

loss of the N=O double bond. The most abundant

compound found in deeply ozonated samples, P3,

presented an accurate mass of 199.0824 Da (C7H11N4O3,

-0.84 ppm) and a fragment ion at mass 142.0609 Da,

corresponding to the loss of CONCH3 groups of the six

member ring. This compound, dessignated 6-amino-5-

(N-formylmethylamino)-1,3.dimethyluracil, had already

been identified as a biological metabolite of caffeine

(Schrader et al., 1999). Coeluting at the same retention

time as P3, compound P4 at m/z 167.0426 (C5H8N2O3Na,

-0.84 ppm) was also detected. As the rest of the

transformation product identified, P4 still conserves the

six member ring structure but a reduction in the DBE

(2.5) suggests the possible disappearance of the double

bond and subsequent hydroxylation on C5 or C6.

In ozonation experiments at pH 8, the identified

transformation products were assumed to correspond to

the oxidation of P2. The accurate mass of 210.0485 Da

was assigned to the sodium adduct of P5, C6H9N3O4Na,

with an error of 0.60 ppm, and tentatively attributed to

the structural formula shown in Fig. 5. Typical neutral

loss of H2O from the carboxylic acid group was also

observed at m/z 170.0560. P6, with experimental mass

158.0560 and without significant fragments was assigned

Chemosphere, 74, 825-831, 2009

to C5H8N3O3, the protonated form of N-methyl-N’-

aminomethylparabamic acid, related to the N,N’-

dimethylparabamic acid identified by Kolonko et al.

(1979) as a by-product of the ozonation of caffeine. P7

was identified with an accurate mass of 227.0774 as

(C8H11N4O4, -0.35 error) a hydroxylated form of the

1,3,7-trimethyluric acid already identified by Telo et al.

(1997). Its fragment ions (170.0561 and 142.0613)

correspond to the opening of the imidazole ring and

losses of the methyl groups.

4. Conclusions

The ozonation of caffeine was specially rapid during the

initial 15 s reaction period, thereafter declining at a much

slower rate. This behaviour suggested the existence of an

intermediate compound whose oxidation was the rate

controlling step. Second order kinetic constants increased

from 0.25 to 1.05 M-1s-1 for increasing pH in the 3-10

range. The amount of ozone consumed per mole of

ozonated caffeine was greater at higher pH values, as was

to be expected from the enhanced ozone decomposition

at basic pH. As for ozone decomposition, it was also

shown to be enhanced when higher concentrations of

caffeine were used. This effect could be attributed to the

presence of a reaction intermediate behaving as a

promoter of ozone decomposition. Irrespective of pH, the

reduction of TOC during ozonation was always low,

indicating the presence of transformation products. A

study of the transformation products identified by LC-

TOF-MS allowed us to propose the two different

degradation routes for pH 3 and 8, as well as to identify

persistent by-products. The results provided evidence that

the ozonation of caffeine followed different mechanistic

routes depending on the ozonation conditions applied.

Most transformation products were the result of the

opening of the imidazole ring after N7C8 double bond of

caffeine was broken.

Acknowledgements

This work has been supported by the Spanish Ministry of

Education (Contracts CTM2005-03080/TECNO and

CONSOLIDER-INGENIO 2010 CSD2006-00044) and

the Dirección General de Universidades e Investigación

de la Comunidad de Madrid under Contract No. PAMB-

000395-0505.

References

Beltran, F.J., 2004, Ozone Reaction Kinetics for Water and

Wastewater Systems, CRC Press LLC, Florida, 2004,

pp. 16-17.

Carballa, M., Omil, F., Lema, J.M., Llompart, M., García-

Jares, C., Rodríguez, I., Gómez, M., Ternes, T., 2004,

Behaviour of pharmaceuticals, cosmetics and hormones

in a sewage treatment plant. Water Res., 38, 2918-2926.

Carr, S.A., Baird, R.B., 2000, Mineralization as a mechanism

for TOC removal: study of ozone/ozone-peroxide

oxidation using FT-IR, Water Res., 34, 4036-4048.

Cataldo, F., 2001, The action of ozone on polymers having

unconjugated and cross- or linearly conjugated

unsaturation: chemistry and technological aspects,

Polym. Degrad. Stab., 73, 511-520.

Dalmázio, I., Santos, L.S., Lopes, R.P., Eberlin, M.N., Augusti,

R., 2005, Advanced oxidation of caffeine in water: on-line

and real-time monitoring by electrospray ionization mass

spectrometry, Environ. Sci. Technol., 39, 5982-5988.

Elovitz, M.S., von Gunten, U., 1999, Hydroxyl radical/ozone

ratios during ozonation processes I. The Rct concept,

Ozone Sci. Eng., 21, 239–260.

Elovitz, M.S., von Gunten, U., Kaiser, H.P., 2000, Hydroxyl

radical/ozone ratios during ozonation processes II. The

effect of temperature, pH, alkalinity and DOM

properties, Ozone Sci. Eng., 22, 123–150.

Fernandez-Alba, A.R., Hernando, D., Agüera, A., Cáceres, J.,

Malato, S., 2002, Toxicity assays: a way for evaluating

AOPs efficiency, Water Res., 36, 4255–4262.

Franke S, Hildebrandt S, Schwarzbauer J, Link M, Francke W.,

1995, Organic compounds as contaminants of the Elbe

river and its tributaries Part II: GCyMS screening for

contaminants of the Elbe water, Fresenius J. Anal.

Chem., 353, 39–49.

Gagné, F., Blaise, C., André, C., 2006, Occurrence of

pharmaceutical products in a municipal effluent and

toxicity to rainbow trout (Oncorhynchus mykiss)

hepatocytes, Ecotox. Environ. Saf., 64, 329-336.

Gossett, R.W., Brown, D.A., Young , D.R., 1983, Predicting

the bioaccumulation of organic compounds in marine

organisms using octanol/water partition coefficients,

Mar. Pollut. Bull., 14, 387–392.

Hoigné, J., 1998, Chemistry of aqueous ozone and

transformation of pollutants by ozone and advanced

oxidation processes, in The Handbook of Environmental

Chemistry, Vol. 5, Part C, Quality and Treatment of

Drinking Water II, Springer, Berlin-Heidelberg.

Kolonko, K. J., Shapiro, R.H., Barkley, R.M., Sievers, R.E.,

1979, Ozonation of caffeine in aqueous solution, J. Org.

Chem., 44, 3769-3778.

Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M.,

Zaugg, S.D., Barber, L.B., Buxton, H.T., 2002,

Pharmaceuticals, hormones, and other organic wastewater

contaminants in U.S. streams, 1999-2000: a national

reconnaissance, Environ. Sci. Technol.,36, 1202-1211.

Kümmerer, K., 2001, Drugs in the environment: emission of

drugs, diagnostic aids and disinfectants into wastewater

by hospitals in relation to other sources – a review.

Chemosphere, 45, 957-969.

Lee, C.S., Rasmussen, T.J., 2006, Occurrence of organic

wastewater compounds in effluent-dominated streams in

Northeastern Kansas, Sci. Total Environ., 371, 258–269.

Petalaa, M., Tsiridisa, V., Samarasb, P. Zouboulisc, A.,

Sakellaropoulosa, G.P., 2006, Wastewater reclamation

by advanced treatment of secondary effluents,

Desalination, 195, 109–118.

Rodríguez, A., Rosal, R., Perdigón, J.A., Mezcua, M., Agüera,

A., Hernando, M.D., Letón, P., Fernández-Alba, A.R.,

García-Calvo, E., 2008 (in press), Ozone-based

Technologies in Water and Wastewater Treatment in D.

Barceló and Mira Petrovic (Eds.) Emerging Contaminants

from Industrial and Municipal Waste (The Handbook of

Environmental Chemistry), Springer-Verlag, Berlín.

Rosal, R. Rodríguez, A, Perdigón-Melón, J.A., Mezcua, M.,

Hernando, M.D., Letón, P., García-Calvo, E., Agüera, A.,

Fernández-Alba, A.R., in press, Removal of

pharmaceuticals and kinetics of mineralization by O3/H2O2

Chemosphere, 74, 825-831, 2009

in a biotreated municipal wastewater, Water Res., in press,

doi: 10.1016/j.watres.2008.06.008.

Schrader, E., Klaunicka, G., Jorritsmaa, U., Neuratha, H.,

Hirsch, K., Kahl , G., Foth, H., 1999, High-performance

liquid chromatographic method for simultaneous

determination of [1-methyl- 14C]caffeine and its eight

major metabolites in rat urine, J. Chrom. B, 726, 195–

201.

Seiler, R.L., Zaugg, S.D., Thomas, J.M., Howcroft, D.L., 1995,

Caffeine and phamaceuticals indicators of water

contamination in wells, Groundwater, 37, 405-410.

Sriamornsak, P., Kennedy, R.A., 2007, Effect of drug

solubility on release behavior of calcium polysaccharide

gel-coated pellets, Eur. J. Pharm,. Sci., 32, 231–239.

Stadler, R.H., Richoz, J., Turezki, R.J., Welti, D.H., Fay, L.B.,

1996, Oxidation of caffeine and related methylxanthines

in ascorbate and polyphenol-driven Fenton-type

oxidations, Free Radical Res., 24, 225-240.

Tang-Liu, D., Williams R., Riegelman, S., 1983, Disposition of

caffeine and its metabolites in man, J. Pharmacol. Exp.

Ther. 24, 180–185.

Tauxe-Wuersch, A., De Alencastro, L.F., Grandjean, D.,

Tarradellas, J., 2005, Occurrence of several acidic drugs

in sewage treatment plants in Switzerland and risk

assessment, Water Res., 39, 1761-1772.

Telo, J.P., Vieira, J.S.C., 1997, Mechanism of free radical

oxidation of caffeine in aqueous solution, J. Chem. Soc.,

Perkin Trans. 2, 1755-1757.

Ternes, T., Bonerz, M., Schmidt, T., 2001, Determination of

neutral pharmaceuticals in wastewater and rivers by liquid

chromatography-electrospray tandem mass spectrometry,

J. Chromatogr. A, 938, 175-185.

Tomiyasu, H., Fukutomi, H., Gordon, G., 1985, Kinetics and

mechanism of ozone decomposition in basic solution,

Inorg. Chem., 24, 2962-2966.

Table 1. Relative ozone consumption and caffeine conversion (xA) during the first 15 s of

ozonation, Second order reaction constant of caffeine (kR) and first order rate constant for

the decomposition of ozone in pure water (kdW).

pH ∆NO3/∆NA (15 s) kR (M-1 s-1) xA (15 s) kdw (s-1)

3 0.49 0.251 0.61 4.1 x 10-4

5 0.94 0.306 0.40 8.3 x 10-4

8 1.23 0.824 0.29 6.1 x 10-3

10 3.76 1.05 0.11 3.2 x 10-2

Table 2. Accurate mass measurements obtained by LC-TOF-MS for Caffeine (A) and its identified

ozonation products at pH 3 and 8.

Compound Formula Experimental Calculated Error (ppm) DBE

mass (m/z) mass (m/z)

pH = 3

A C8H11N4O2 195.0876 195.0875 0.27 5.5

P1 C8H11N4O5 243.0724 243.0723 0.02 5.5

C8H10N4O5Na 265.0547 265.0543 1.35 5.5

C8H9N4O4 225.0620 225.0618 0.75 6.5

P2 C8H12N4O4Na 251.0751 251.0750 0.09 4.5

C7H10N3O3 184.0713 184.0716 -1.99 4.5

C5H7N2O2 127.0499 127.0502 -2.37 3.5

P3 C7H11N4O3 199.0824 199.0825 -0.84 4.5

C5H8N3O2 142.0609 142.0611 -1.43 3.5

P4 C5H8N2O3Na 167.0426 167.0427 -0.67 2.5

C3H9N2O 89.0713 89.0709 4.04 0.5

pH = 8

A C8H11N4O2 195.0876 195.0875 -0.78 5.5

P5 C6H9N3O4Na 210.0485 210.0484 0.60 3.5

C6H8N3O3 170.0560 170.0560 -0.10 4.5

P6 C5H8N3O3 158.0560 158.0558 -1.37 3.5

P7 C8H11N4O4 227.0774 227.0774 -0.35 5.5

C6H8N3O3 170.0561 170.0560 0.48 4.5

C5H8N3O2 142.0613 142.0611 1.38 3.5

Fig. 1. Evolution of the concentration of caffeine for runs performed at pH 3 (■), 5 (▲) and 10 (●) . Solid

lines and right scale correspond to the ozone profile for the same runs.

0.0

0.2

0.4

0.6

0.8

1.0

0 200 400 600 800 1000

t (s)

CA/C

A0

0.0

0.1

0.1

0.2

0.2

0.3

0.3

CO

3 (

mM

)

0

0.2

0.4

0.6

0.8

1

1.2

1.4

0 0.2 0.4 0.6 0.8

∫CO3dt (mM s)

ln(C

Ao/C

A)

Fig. 2. Ozone exposure plot for the ozonation of caffeine for different ozone-to-caffeine initial molar ratios.

pH 3 with molar ratios 1.07 (□) and 0.70 (■); pH 5 with ratios 1.01 (∆) and 0.79 (▲), pH 8 and ratios 0.94

(�) and 0.78 (�) and pH 10 at ratios 1.05 (□) and 0.75 (●).

Fig. 3. Concentration of caffeine (■) and TOC (□) during an ozonation run with continuous flow of ozone at

pH 3, 25ºC and CO3(g) = 12 g/Nm3. Solid line and right scale show the concentration of dissolved ozone.

0

10

20

30

40

50

60

0 2000 4000 6000

t (s)

CA, T

OC

(p

pm

)

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

CO

3 (

pp

m)

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

CAo (mM)

kd (

s-1

)

Fig. 4. Apparent first-order constant for the decomposition of ozone during the first part of runs at pH 3 (□),

pH 5 (∆) and pH 8 (□).

a)

b)

Fig. 5. Main products identified by LC-TOF-MS and proposed pathway for the ozonation of caffeine a) in

acidic media at pH 3 and b) under alkaline conditions at pH 8.